High Precise Localization of Mobile Robot by Three Times Pose Correction

Abstract

With the rapid development of artificial intelligence, lots of simple and repetitive tasks can be gradually taken over by mobile robots, such as power inspection, factory pickup, and so forth. In actual operation, mobile robots are required to equip with a high precise localization system. In this paper, we fuse information of the odometry, inertial measure unit (IMU), and laser radar to reduce localization error by combining three times pose correction from the Extended Kalman Filter (EKF), the adaptive Monte Carlo localization (AMCL), and the enhanced scan matching algorithm. The ensemble algorithm can overcome the difficulty of localization when robot run in the harsh environment where is not suitable for using GPS, vision, and Ultra-Wideband (UWB). The experimental results show that, compared with a single method, the hybrid localization method can make the robot achieve high precise localization in both indoor and outdoor environments.